<?xml version="1.0" encoding="utf-8"?>
<TEI xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:hal="http://hal.archives-ouvertes.fr/" xmlns:gml="http://www.opengis.net/gml/3.3/" xmlns:gmlce="http://www.opengis.net/gml/3.3/ce" version="1.1" xsi:schemaLocation="http://www.tei-c.org/ns/1.0 http://api.archives-ouvertes.fr/documents/aofr-sword.xsd">
  <teiHeader>
    <fileDesc>
      <titleStmt>
        <title>HAL TEI export of hal-04222825</title>
      </titleStmt>
      <publicationStmt>
        <distributor>CCSD</distributor>
        <availability status="restricted">
          <licence target="https://creativecommons.org/publicdomain/zero/1.0/">CC0 1.0 - Universal</licence>
        </availability>
        <date when="2026-05-23T02:27:09+02:00"/>
      </publicationStmt>
      <sourceDesc>
        <p part="N">HAL API Platform</p>
      </sourceDesc>
    </fileDesc>
  </teiHeader>
  <text>
    <body>
      <listBibl>
        <biblFull>
          <titleStmt>
            <title xml:lang="en">Batch-less stochastic gradient descent for compressive learning of deep regularization for image denoising</title>
            <author role="aut">
              <persName>
                <forename type="first">Hui</forename>
                <surname>Shi</surname>
              </persName>
              <email type="md5">dec62d2ef9b4b1c30b51f4e61e24299f</email>
              <email type="domain">u-bordeaux.fr</email>
              <idno type="idhal" notation="numeric">1089606</idno>
              <idno type="halauthorid" notation="string">964292-1089606</idno>
              <affiliation ref="#struct-27730"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Yann</forename>
                <surname>Traonmilin</surname>
              </persName>
              <email type="md5">4831cf36addf5248bbf5a70552f5dc29</email>
              <email type="domain">math.u-bordeaux.fr</email>
              <idno type="idhal" notation="numeric">1077290</idno>
              <idno type="halauthorid" notation="string">2547153-1077290</idno>
              <affiliation ref="#struct-27730"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">J-F</forename>
                <surname>Aujol</surname>
              </persName>
              <email type="md5">412dfc4a768d38415e63c3ad7010da51</email>
              <email type="domain">math.u-bordeaux.fr</email>
              <idno type="idhal" notation="string">jaujol</idno>
              <idno type="idhal" notation="numeric">184062</idno>
              <idno type="halauthorid" notation="string">1363689-184062</idno>
              <idno type="ORCID">https://orcid.org/0000-0001-6716-0509</idno>
              <idno type="IDREF">https://www.idref.fr/081164432</idno>
              <affiliation ref="#struct-27730"/>
            </author>
            <editor role="depositor">
              <persName>
                <forename>Yann</forename>
                <surname>Traonmilin</surname>
              </persName>
              <email type="md5">4831cf36addf5248bbf5a70552f5dc29</email>
              <email type="domain">math.u-bordeaux.fr</email>
            </editor>
            <funder ref="#projanr-51031"/>
          </titleStmt>
          <editionStmt>
            <edition n="v1" type="current">
              <date type="whenSubmitted">2023-09-29 14:02:31</date>
              <date type="whenModified">2026-02-07 05:26:10</date>
              <date type="whenReleased">2023-10-02 13:46:45</date>
              <date type="whenProduced">2024-03-13</date>
              <date type="whenEndEmbargoed">2023-09-29</date>
              <ref type="file" target="https://hal.science/hal-04222825v1/document">
                <date notBefore="2023-09-29"/>
              </ref>
              <ref type="file" subtype="author" n="1" target="https://hal.science/hal-04222825v1/file/deep_reg_sketch.pdf" id="file-4222825-3682979">
                <date notBefore="2023-09-29"/>
              </ref>
              <ref type="externalLink" target="http://arxiv.org/pdf/2310.03085"/>
            </edition>
            <respStmt>
              <resp>contributor</resp>
              <name key="178764">
                <persName>
                  <forename>Yann</forename>
                  <surname>Traonmilin</surname>
                </persName>
                <email type="md5">4831cf36addf5248bbf5a70552f5dc29</email>
                <email type="domain">math.u-bordeaux.fr</email>
              </name>
            </respStmt>
          </editionStmt>
          <publicationStmt>
            <distributor>CCSD</distributor>
            <idno type="halId">hal-04222825</idno>
            <idno type="halUri">https://hal.science/hal-04222825</idno>
            <idno type="halBibtex">shi:hal-04222825</idno>
            <idno type="halRefHtml">&lt;i&gt;Journal of Mathematical Imaging and Vision&lt;/i&gt;, 2024</idno>
            <idno type="halRef">Journal of Mathematical Imaging and Vision, 2024</idno>
            <availability status="restricted">
              <licence target="https://about.hal.science/hal-authorisation-v1/">HAL Authorization<ref corresp="#file-4222825-3682979"/></licence>
            </availability>
          </publicationStmt>
          <seriesStmt>
            <idno type="stamp" n="CNRS">CNRS - Centre national de la recherche scientifique</idno>
            <idno type="stamp" n="IMB">Institut de Mathématiques de Bordeaux</idno>
            <idno type="stamp" n="INSMI">CNRS-INSMI - INstitut des Sciences Mathématiques et de leurs Interactions</idno>
            <idno type="stamp" n="ANR">ANR</idno>
            <idno type="stamp" n="UNIVERSITE-BORDEAUX">Université de Bordeaux</idno>
          </seriesStmt>
          <notesStmt>
            <note type="audience" n="2">International</note>
            <note type="popular" n="0">No</note>
            <note type="peer" n="1">Yes</note>
          </notesStmt>
          <sourceDesc>
            <biblStruct>
              <analytic>
                <title xml:lang="en">Batch-less stochastic gradient descent for compressive learning of deep regularization for image denoising</title>
                <author role="aut">
                  <persName>
                    <forename type="first">Hui</forename>
                    <surname>Shi</surname>
                  </persName>
                  <email type="md5">dec62d2ef9b4b1c30b51f4e61e24299f</email>
                  <email type="domain">u-bordeaux.fr</email>
                  <idno type="idhal" notation="numeric">1089606</idno>
                  <idno type="halauthorid" notation="string">964292-1089606</idno>
                  <affiliation ref="#struct-27730"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Yann</forename>
                    <surname>Traonmilin</surname>
                  </persName>
                  <email type="md5">4831cf36addf5248bbf5a70552f5dc29</email>
                  <email type="domain">math.u-bordeaux.fr</email>
                  <idno type="idhal" notation="numeric">1077290</idno>
                  <idno type="halauthorid" notation="string">2547153-1077290</idno>
                  <affiliation ref="#struct-27730"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">J-F</forename>
                    <surname>Aujol</surname>
                  </persName>
                  <email type="md5">412dfc4a768d38415e63c3ad7010da51</email>
                  <email type="domain">math.u-bordeaux.fr</email>
                  <idno type="idhal" notation="string">jaujol</idno>
                  <idno type="idhal" notation="numeric">184062</idno>
                  <idno type="halauthorid" notation="string">1363689-184062</idno>
                  <idno type="ORCID">https://orcid.org/0000-0001-6716-0509</idno>
                  <idno type="IDREF">https://www.idref.fr/081164432</idno>
                  <affiliation ref="#struct-27730"/>
                </author>
              </analytic>
              <monogr>
                <idno type="halJournalId" status="VALID">15761</idno>
                <idno type="issn">0924-9907</idno>
                <idno type="eissn">1573-7683</idno>
                <title level="j">Journal of Mathematical Imaging and Vision</title>
                <imprint>
                  <publisher>Springer Verlag</publisher>
                  <date type="datePub">2024-03-13</date>
                </imprint>
              </monogr>
              <idno type="arxiv">2310.03085</idno>
            </biblStruct>
          </sourceDesc>
          <profileDesc>
            <langUsage>
              <language ident="en">English</language>
            </langUsage>
            <textClass>
              <keywords scheme="author">
                <term xml:lang="en">Regularization</term>
                <term xml:lang="en">Compressive learning</term>
                <term xml:lang="en">Denoising</term>
                <term xml:lang="en">deep priors</term>
              </keywords>
              <classCode scheme="halDomain" n="info.info-ai">Computer Science [cs]/Artificial Intelligence [cs.AI]</classCode>
              <classCode scheme="halDomain" n="info.info-it">Computer Science [cs]/Information Theory [cs.IT]</classCode>
              <classCode scheme="halTypology" n="ART">Journal articles</classCode>
              <classCode scheme="halOldTypology" n="ART">Journal articles</classCode>
              <classCode scheme="halTreeTypology" n="ART">Journal articles</classCode>
            </textClass>
            <abstract xml:lang="en">
              <p>We consider the problem of denoising  with the help of prior information taken from a database of clean signals or images. Denoising with variational methods is very efficient if a regularizer well adapted to the nature of the data is available. Thanks to the maximum a posteriori Bayesian framework, such regularizer can be systematically linked with the distribution of the data. With deep neural networks (DNN), complex distributions can be recovered from a large training database.To reduce the computational burden of this task,  we  adapt the compressive learning framework to the learning of regularizers parametrized by DNN. We propose two variants of stochastic gradient descent (SGD) for the recovery of deep regularization parameters from a heavily compressed database. These algorithms outperform the initially proposed method that was limited to low-dimensional signals, each iteration using information from the whole database. They also benefit from classical SGD convergence guarantees. Thanks to these improvements we show that this method can be applied for patch based image denoising.}</p>
            </abstract>
          </profileDesc>
        </biblFull>
      </listBibl>
    </body>
    <back>
      <listOrg type="structures">
        <org type="laboratory" xml:id="struct-27730" status="VALID">
          <idno type="IdRef">155051342</idno>
          <idno type="ISNI">0000000123024783</idno>
          <idno type="RNSR">200711916B</idno>
          <idno type="ROR">https://ror.org/05m3r1b84</idno>
          <orgName>Institut de Mathématiques de Bordeaux</orgName>
          <orgName type="acronym">IMB</orgName>
          <date type="start">2003-01-01</date>
          <desc>
            <address>
              <addrLine>351 cours de la Libération 33405 TALENCE CEDEX</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.math.u-bordeaux.fr/IMB/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-259761" type="direct"/>
            <relation active="#struct-409745" type="direct"/>
            <relation name="UMR5251" active="#struct-441569" type="direct"/>
          </listRelation>
        </org>
        <org type="institution" xml:id="struct-259761" status="VALID">
          <idno type="IdRef">175206562</idno>
          <idno type="ROR">https://ror.org/057qpr032</idno>
          <orgName>Université de Bordeaux</orgName>
          <orgName type="acronym">UB</orgName>
          <desc>
            <address>
              <addrLine>35, place Pey Berland - 33076 Bordeaux</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.u-bordeaux.fr/</ref>
          </desc>
        </org>
        <org type="institution" xml:id="struct-409745" status="VALID">
          <orgName>Institut Polytechnique de Bordeaux</orgName>
          <orgName type="acronym">Bordeaux INP</orgName>
          <desc>
            <address>
              <addrLine>Avenue des facultés33405 Talence</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">https://www.bordeaux-inp.fr/fr</ref>
          </desc>
        </org>
        <org type="regroupinstitution" xml:id="struct-441569" status="VALID">
          <idno type="IdRef">02636817X</idno>
          <idno type="ISNI">0000000122597504</idno>
          <idno type="ROR">https://ror.org/02feahw73</idno>
          <orgName>Centre National de la Recherche Scientifique</orgName>
          <orgName type="acronym">CNRS</orgName>
          <date type="start">1939-10-19</date>
          <desc>
            <address>
              <country key="FR"/>
            </address>
            <ref type="url">https://www.cnrs.fr/</ref>
          </desc>
        </org>
      </listOrg>
      <listOrg type="projects">
        <org type="anrProject" xml:id="projanr-51031" status="VALID">
          <idno type="anr">ANR-20-CE40-0001</idno>
          <orgName>EFFIREG</orgName>
          <desc>Régularisation performante de problèmes inverses en grande dimension pour le traitement de données</desc>
          <date type="start">2020</date>
        </org>
      </listOrg>
    </back>
  </text>
</TEI>